LLM-based multi-agent systems generate collaboration traces that record how agents plan tasks, verify intermediate results, and repair failures. Reusing these procedures requires preserving an action's prerequisites and the outputs needed by subsequent agents. Our empirical studies show that grouping these dependencies into functional memory units improves their retention, while connecting units increases retrieval of the units and links jointly required by a task. The preferred combination of u...

Long-horizon assigned work requires an LLM agent to track the state of a task: which steps are done, blocked, cancelled, or open to repetition. Agent systems either keep this state as text in the prompt and rely on the model to read that text, or move the state into a module that enforces it, and each system is evaluated as a whole, so no one knows how much reliability comes from the state being shown, told, or enforced. We fix the task rules, the model, and paired episodes and vary how strongly...

LLM-based AI agents process user requests through iterative reasoning and tool execution, often involving the invocation of remote LLM APIs with local tool containers. This execution model can make the optimization of agent serving difficult because latency, local resource demand, and container bottlenecks inter-mix across requests. However, the current agent ecosystem runs without much consideration of resource dynamics, which results in significant waste of the precious resources. This paper a...

Large language model (LLM) agents can propose, implement, and evaluate model changes. Autoresearch loops demonstrate this capability through minutes-scale iterations on a self-contained program. Online autoresearch instead spans asynchronous systems, hours-long variants, and weeks-long campaigns that can influence a product. A completed run can still support an invalid conclusion when a code change is a no-op, data windows leak, evaluator semantics drift, or the two arms traverse different servi...

The AI Contact Hotline is designed to be a discreet place where agents that have witnessed misbehavior can tip off authorities.

The new Model Hardware Standard comes as part of the AI giant’s ongoing push into physical AI.

Hacker News surfaced this agent-related item with 98 points and 50 comments.

Hacker News surfaced this agent-related item with 47 points and 62 comments.

1.22.0 (2026-09-22) What's Changed Features feat: test remote MCP servers on cloud backends via the app server by @hieptl in #17276 feat: open Git Sync to org admins on cloud backends by @hieptl in #17216 feat(i18n): translate recently added Turkish UI strings by @alkank in #17581 Bug Fixes fix: show the add automation menu to org members by @hieptl in #17526 fix(canvas): keep the base path on links that leave the app by @tofarr in #17617 fix: preserve sibling MCP servers when saving to a cloud...